CardioVascular disease is one of the most common diseases in the world. The aim of this study is to support the diagnosis of the Heart disease using a simple and efficient classification system based on Support Vector Machine (SVM) classifier. The proposed method contains three stages (1) Feature selection based on Predictive Scores, (2) Dimensionality Reduction of the data set using Singular Value Decomposition (SVD) and (3) Classification of the processed dataset using Nonlinear SVM with Radial Basis Function (RBF) Kernel. The Statlog and Cleveland Heart disease data sets are taken from the UCI learning data repository for experiments. A maximum accuracy of 94.20% was achieved according to 10-fold cross-validation technique. A third dataset, which is a combination of these two data sets was prepared. The proposed method achieved good results when applied in combination of these datasets, despite the increasing number of patients. The results demonstrate that the proposed method performed better when compared with the experimental results of the previously reported classification methods by researchers. This study might help Physicians to identify the patients with Heart disease at early stages so that disease can be cured.

Heart Disease, Predictive Scores based feature selection, Singular Value Decomposition (SVD), Support Vector Machine (SVM), Radia Basis Kernel Function (RBF)